This paper proposes a data mining approach to modeling relationships among categories in image collection. In our approach, with image feature grouping, a visual dictionary is cre...
Methods for ranking World Wide Web resources according to their position in the link structure of the Web are receiving considerable attention, because they provide the first e...
In this paper, we describe a methodology to estimate the geographic coverage of the web without the need for secondary knowledge or complex geo-tagging. This is achieved by random...
Robert Pasley, Paul Clough, Ross S. Purves, Floria...
The Rich News system, that can automatically annotate radio and television news with the aid of resources retrieved from the World Wide Web, is described. Automatic speech recogni...
Mike Dowman, Valentin Tablan, Hamish Cunningham, B...
In this paper, we address the question of what kind of knowledge is generally transferable from unlabeled text. We suggest and analyze the semantic correlation of words as a gener...